U.S. patent application number 12/547580 was filed with the patent office on 2011-03-03 for system and method for providing network resiliency.
This patent application is currently assigned to Noblis, Inc.. Invention is credited to David Anthony Garbin, Joseph Edward Knepley, JR..
Application Number | 20110051594 12/547580 |
Document ID | / |
Family ID | 43624754 |
Filed Date | 2011-03-03 |
United States Patent
Application |
20110051594 |
Kind Code |
A1 |
Knepley, JR.; Joseph Edward ;
et al. |
March 3, 2011 |
System and Method for Providing Network Resiliency
Abstract
The invention is directed toward an automated system and process
to provide resiliency to a network infrastructure where traditional
established methods tend to fail. The system analyzes the
underlying physical characteristics of the network topology to
identify physically diverse backup routing paths among nodes in the
network. The system takes as input the number of diverse paths
desired and a characteristic(s) used to choose the most desirable
paths, such as delay, flow, or cost. The invention searches for
multiple solutions for the selected number of diverse paths within
the network.
Inventors: |
Knepley, JR.; Joseph Edward;
(Falls Church, VA) ; Garbin; David Anthony;
(Reston, VA) |
Assignee: |
Noblis, Inc.
|
Family ID: |
43624754 |
Appl. No.: |
12/547580 |
Filed: |
August 26, 2009 |
Current U.S.
Class: |
370/216 |
Current CPC
Class: |
H04L 45/126
20130101 |
Class at
Publication: |
370/216 |
International
Class: |
G06F 11/00 20060101
G06F011/00 |
Claims
1. A method for providing resiliency in a network comprising:
identifying a network topology including a plurality of nodes and
links, each link connecting two nodes; receiving criteria for a
route through the network between a source node and a destination
node, the criteria including a number of diverse paths connecting
the source node and the destination node, a path diversity
criteria, and an optimization parameter for use in ranking paths;
finding paths through the network using a compass direction by:
locating a path in the network from the source node to the
destination node by: beginning at the source node, following a link
most closely aligned with the selected compass direction from the
source node to a next node; extending the path from the next node
by following a link most closely aligned with the selected compass
direction to another node; repeating the extending step until the
path reaches the destination node; repeating the locating step,
while omitting links that are part of a previously located path, to
determine additional diverse paths from the source node to the
destination node; applying the optimization parameter to rank the
located paths to identify a primary path and one or more backup
paths for the selected compass direction; repeating the finding
step for each compass direction; for each compass direction that
found the number of diverse paths specified by the route criteria,
applying the optimization parameter to rank the primary paths to
determine one or more optimal paths from the source node to the
destination node; and outputting the one or more optimal paths.
2. The method of claim 1, wherein the path diversity criteria is
link diversity, such that each located path cannot share any
links.
3. The method of claim 1, wherein the path diversity criteria is
link-and-node diversity, such that each located path cannot share
any node or link, and, wherein nodes that are part of a previously
located path are further omitted when the locating step is
repeated.
4. The method of claim 1, wherein the optimization criteria is at
least one of distance, cost, availability, bandwidth, throughput,
capacity, delay, and speed.
5. The method of claim 1, wherein the compass directions are north,
south, east, and west.
6. A method for providing resiliency in a network comprising:
identifying a network topology including a plurality of nodes and
links, each link connecting two nodes; receiving criteria for a
route through the network between a source node and a destination
node; finding paths in a selected compass direction by: locating a
path in the network from the source node to the destination node by
forcing the path in the selected compass direction; repeating the
locating step as long as it is possible to determine additional
diverse paths from the source node to the destination node while
omitting any link that is part of a previously located path;
ranking the located paths according to the criteria to identify a
primary path and one or more backup paths in the selected compass
direction; repeating the finding step for each compass direction;
ranking the primary path of each compass direction to determine the
optimal compass direction; and selecting the paths corresponding to
the optimal compass direction as the routes through the
network.
7. The method of claim 6, wherein the criteria includes at least
one of availability, cost, distance, speed, throughput, bandwidth,
capacity, and delay.
8. The method of claim 6, wherein the route criteria includes a
number of diverse paths connecting the source node and the
destination node.
9. The method of claim 6, wherein the route criteria includes a
path diversity criteria.
10. The method of claim 9, wherein the path diversity criteria is
link diversity, such that each determined path cannot share any
links.
11. The method of claim 9, wherein the path diversity criteria is
link-and-node diversity, such that each determined path cannot
share any node or link, and, wherein nodes that are part of a
previously located path are further omitted when the locating step
is repeated.
12. A computer system for providing resiliency in a network
comprising: an input device for receiving a network topology
including a plurality of nodes and links, each link connecting two
nodes, and for receiving criteria for a route through the network
between a source node and a destination node; a network path
processor for finding paths in a direction by: locating a path in
the network from the source node to the destination node by forcing
the path in the selected direction; repeating the locating step
while it is possible to determine additional diverse paths from the
source node to the destination node, without using any link that is
part of a previously located path; the network path processor
repeating the finding step for a least one other direction; an
output device for outputting the paths associated with each
direction.
13. The system of claim 12, wherein the route criteria includes at
least one of availability, cost, distance, speed, throughput,
bandwidth, capacity, and delay.
14. The system of claim 12, wherein the network path processor
further applies the route criteria to rank the paths associated
with each direction.
15. The system of claim 12, wherein the route criteria includes a
number of diverse paths connecting the source node and the
destination node.
16. The system of claim 12, wherein the finding further comprises:
ranking the located paths using the route criteria to identify a
primary path and one or more backup paths for the selected
direction.
17. The system of claim 12, wherein the route criteria specifies
link diversity, such that each determined path cannot share any
links.
18. The system of claim 12, wherein the route criteria specifies
link-and-node diversity, such that each determined path cannot
share a node or link, and wherein nodes that are part of a
previously located path are further omitted when locating is
repeated.
Description
FIELD OF THE INVENTION
[0001] The disclosed subject matter relates to the field of
networking, and more particularly to methods and systems for
providing resiliency in a network based on selected paths between
nodes in the network.
BACKGROUND OF THE INVENTION
[0002] A network is a group of objects, or nodes, connected to each
other by links. Data, people, or supplies may be transported
between the nodes in a network via the links. Examples of common
networks include train stations connected by tracks, a community
connected by roads, a communications network connected by wires,
fiber optics or satellite links, a supply network connected by
delivery routes, and logic gates connected by paths in an
integrated circuit. In some networks, such as a ring network, there
is only one path between nodes. In others, such as a mesh network,
there may be several different paths between any two nodes in the
network.
[0003] Many networks provide critical functions to business and
governments. Should there be an interruption in a supply network, a
business may fail to meet its contract obligations. A failure in a
communications network may mean critical messages, such as those
needed for directing air traffic, are lost. In these types of
networks, it is desirable to have a mesh network with backup paths
to connect nodes should a primary path fail. This is known as
network resiliency. Resiliency ensures that communication between
nodes is not interrupted by the failure of any one link between
nodes. In some networks, uninterrupted communication is so
important that more than one backup path is often desired to
increase network resiliency.
[0004] Alternate routes, or backup paths, may be identified using
methods such as the "shortest path" method, which calculates a
diverse path between nodes in a network by minimizing the number of
links first and total link distance between nodes second. However,
traditional methods are not reliable when more than one backup path
is desired. For example, in a network where desired network
resiliency requires three diverse paths among nodes, traditional
methods fail to calculate the desired number of paths up to 5% of
the time. In many networks the failure to provide at least three
diverse paths results in an unacceptable failure rate. Therefore
there exists a need for a system that will identify multiple
alternate paths for all node pairs in a network where traditional
methods fail.
SUMMARY OF THE INVENTION
[0005] Embodiments consistent with the invention provide methods
and systems for providing resiliency in a network. Resiliency is
provided by locating multiple diverse paths in a selected direction
between a source node and a destination node. This is accomplished
by forcing an initial path from the source node to the destination
node in the selected direction, for example the northernmost path.
Additional diverse paths are found by removing the links used in
previously found paths, and forcing the path once again in the
selected direction. When a path from the source node to the
destination node cannot be found without using a link used by a
previously found path, all paths in the selected direction have
been found, and comprise the diverse paths between the source node
and the destination node in the selected direction. Diverse paths
may be found in this way for each desired direction. This forced
directionalized routing allows embodiments of the invention to
provide more diverse paths between a source node and a destination
node than traditional path location tools.
[0006] In addition, embodiments consistent with the invention may
also use path criteria, such as a minimum number of paths or path
optimization criteria, for determining the direction with the most
desirable set of paths between the source node and the destination
node. For example, at least three diverse paths may be desired, and
the direction that contains the fastest route may be more desirable
than the other directions. The path criteria may help select the
direction with the optimal set of diverse paths between the source
node and destination node in the network.
[0007] Embodiments consistent with the invention may also recommend
a change to the network topology to accommodate additional diverse
paths. If a desired number of diverse paths between a source node
and a destination node cannot be located, the invention may suggest
the addition of links and/or nodes to create the desired number of
diverse paths.
[0008] It is to be understood that both the foregoing general
description and the following detailed description are exemplary
and explanatory only and are not restrictive of the invention, as
claimed.
[0009] The accompanying drawings, which are incorporated in and
constitute a part of this specification, illustrate one embodiment
of the invention and together with the description, serve to
explain the principles of the invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] FIG. 1 is an exemplary network topology, consistent with
embodiments of the present invention.
[0011] FIG. 2 is another an exemplary network topology, consistent
with embodiments of the present invention.
[0012] FIG. 3 is a diagram of a system, consistent with embodiments
of the present invention.
[0013] FIG. 4 is a flow diagram illustrating an exemplary network
routing process consistent with embodiments of the present
invention.
[0014] FIG. 5 is a flow diagram of an exemplary directionalized
routing process, consistent with embodiments of the present
invention.
[0015] FIGS. 6a-6n are exemplary network topologies depicting a
directionalized routing processes consistent with embodiments of
the present invention.
DESCRIPTION OF THE EMBODIMENTS
[0016] Reference will now be made in detail to exemplary
embodiments of the invention, examples of which are illustrated in
the accompanying drawings. Wherever possible, the same reference
numbers will be used throughout the drawings to refer to the same
or like parts.
[0017] FIG. 1 is an exemplary network topology, consistent with
embodiments of the present invention. The nodes of this network may
represent, for example, cities or transfer stations. The links of
the network may represent supply lines, train tracks, or fiber
optic or other communication lines. This is an example of a mesh
network.
[0018] FIG. 2 is another exemplary network topology, consistent
with embodiments of the present invention. The nodes of this
network may represent, for example, add/drop multiplexers in a
power network.
[0019] Many networks, such as those shown in FIGS. 1 and 2, provide
critical functions to business, cities, and governments. Network
failures may cause great financial hardship, great inconvenience,
or even threaten physical harm. Therefore, these networks require
resiliency in the form of backup paths between nodes, so that if a
primary path between nodes fails, the flow of data, things, or
people remains un-interrupted. In some cases, the risk that a
backup path may fail must also be accounted for. Traditional
methods for identifying multiple paths between nodes in a network
tend to fail when three or more paths are required between the
nodes. Because of this, there exists a need for a network routing
system that can determine multiple physically diverse backup
paths.
[0020] It is also highly desirable that a system for finding
physically diverse paths be flexible and configurable. Some
networks rely on speed, others value reliability, while still
others favor low cost. It is desirable to take these variables into
consideration when calculating backup paths and determining which
backup paths to select. The burden of calculating backup paths, and
the failure rate of traditional methods, increases when a network
requires the backup paths to have node diversity in addition to
link diversity, or when a network requires three, four, or more
diverse paths.
[0021] Additionally, when the desired number of paths cannot be
found with the existing network topology, there exists a need for a
network routing system to determine where additional nodes or links
may be placed within the network topology to yield at least one
potential solution. For example, some networks require a primary
path and two diverse backup paths, and when the current network
topology will not support this number of backup paths there exists
a need to determine where additional links or nodes can be placed
in the network topology so that all source to destination nodes can
have three diverse paths.
[0022] A network routing system with directionalized routing in
accordance with the invention addresses these needs. The system is
flexible and configurable, having the capability for users to
select n disjoint routes having node or link diversity. It also has
the capability to allow the user to select performance criteria to
be applied to the paths to determine which is the best choice for a
primary path.
[0023] Systems consistent with the present invention analyze the
underlying physical characteristics of network topology to identify
physically diverse backup routing paths among nodes in the network.
An exemplary system may identify paths by looping through all
compass directions to determine all possible solutions for the
specified number of diverse paths. If more than a single solution
for the desired number of paths is found, the solutions may be
ranked by optimization variable(s) to determine the path(s) chosen.
Such an exemplary system enables any given number of backup paths
to be found for each source to destination node pair in a
network.
[0024] FIG. 3 is a diagram of a routing system, consistent with the
embodiments of the present invention. System 300 contains data
representing a network topology 305, which is stored in a memory
storage device. Network topology 305 may be received, for example,
through an input device 345 or from an external storage device 350.
Network topology 305 may be data describing links and nodes that
make up a network. For example, nodes may be identified by
longitude and latitude, associated devices, street intersections,
or IP address. Links may be described using the nodes they connect
to, bandwidth, or track name.
[0025] System 300 also contains instructions 310, bus 335, and
routing processor 320. Instructions 310 may be stored in main
memory and received from either an external or internal computer
readable medium. Instructions 310 may be sent to routing processor
320 via bus 335. Routing processor 320 may execute instructions
310, using network topology 305 and/or routing parameters 315.
Routing parameters 315 may be obtained through input device 345, or
external storage 350. Routing parameters 315 may be, for example,
path endpoints, the number of diverse paths desired between nodes
in the network, path diversity selection, and/or path performance
parameters. The path endpoints may represent the starting node and
the ending node in the network topology. A path diversity selection
may be either link diverse or link-and-node diverse. Link diverse
paths cannot share the same link, but may share the same nodes.
Link-and-node diverse paths cannot share the same links nor the
same nodes.
[0026] Path performance parameters indicate what criteria to use to
rank one path above another path in the network. Minimizing delay
is one example of a path performance parameter. Under this criteria
a path representing a faster travel time between the path endpoints
would rank higher than a path with a slower travel time between the
same two endpoints. Another example of a path performance parameter
is availability. Under this criteria a path with fewer
interruptions in service over time would rank higher than a path
with more interruptions in service. For example, in a network of
train stations and train tracks, track lines requiring frequent
repairs in the past would have a lower availability ranking than
track lines having a scant repair history. Weather delays on a
track line could also affect a line's availability ranking. Another
example of a path performance parameter is cost. Under this
criteria paths having a lower transportation cost would rank higher
than paths with higher transportation costs. An exemplary system
consistent with the present invention is not limited to the three
examples above and may have other path performance parameters.
[0027] Routing processor 320 may determine primary and/or alternate
paths between nodes in the network using instructions 310. The
primary and/or alternate paths 325, may be stored on a computer
readable medium, and/or presented to a user via display 340, etc.
The display may be, for example, a computer monitor connected to
system 300 or a printed report. In certain embodiments, if routing
processor 320 cannot calculate the desired number of paths
requested by routing parameters 315, routing processor 320 may
suggest additional nodes or links 330 be added to network
topology.
[0028] FIG. 4 is a flow diagram illustrating an exemplary network
routing process consistent with embodiments of the present
invention. In certain embodiments, this process may be implemented
according to instructions 310. In an embodiment implemented using
system 300, as shown in FIG. 4, routing processor 320 receives data
describing a network topology (step 400) for a network such as the
one shown in FIG. 6a. It stores the network topology 305 in a
memory storage device. Next, system 300 receives routing parameters
315, including data representing the path endpoints (step 405), and
data describing the desired path characteristics (step 410), such
as the number of paths desired, the path diversity criteria, and
the path performance parameters as described above. Routing
parameters 315 may be stored in a memory storage device, such as
350, by routing processor 320.
[0029] After routing processor 320 receives the desired number of
paths, it determines the degree of network connectivity in step 415
and compares the connectivity to the desired number of paths in
step 420. If the desired number of paths exceeds the network
connectivity (step 420, Yes) then routing processor 320 warns the
user, for example, through display 340 that the current network
topology will not support the requested number of
source-to-destination diverse paths, and may recommend new link(s)
and/or node(s) in step 445. If the network connectivity supports
the desired number of paths, (step 420, No), routing processor 320
may use traditional methods, such as the shortest path, to find the
desired number of paths, as shown in step 425. If this step fails
to find the desired number of paths (step 430, No), routing
processor 320 will execute instructions 310 for the directionalized
network routing method (step 435). If the desired number of paths
are still not found (step 440, No), routing processor 310 will
recommend new links and/or nodes.
[0030] FIG. 5 is a flow diagram of an exemplary directionalized
routing process consistent with embodiments of the present
invention. In an embodiment implemented using system 300, to begin,
routing processor 320 selects a compass direction, for example
North, in step 505. Step 510 begins at the source node, e.g. as
received in step 405. From this node, routing processor 320 may
choose an available link to the next available node that is most
closely aligned with the compass direction selected in step 505, as
shown in step 515. Distance modifications can be used to find a
link most closely aligned with a compass direction. For example,
the northern, eastern, southern and western most links may be
determined by modifying the distance between connected node pairs
from the actual distance to a transformation of the distant end
node's latitude or longitude, using distance modifications such as
those described below: [0031] Northernmost distance from node x1 to
node x2 (x2 is not destination node)
[0031] D(x1,x2)=1.5 {[180-latitude(x2)]/2} [0032] Easternmost
distance from node x1 to node x2 (x2 is not destination. Each "If"
statement is tested in sequence)
[0032] AL(x2)=AdjustedLongitude(x2)=longitude(x2)-longitude(x1)
If (AL(x2)<0) then AL(x2)=AL(x2)+360
If (AL(x2).ltoreq.180) then AL(x2)=180-AL(x2)
If (AL(x2)>180) then AL(x2)=540-AL(x2)
AL(x2)=AL(x2)+1
D(x1,x2)=1.5 [AL(x2)/4] [0033] Southernmost distance from node x1
to node x2 (x2 is not destination node)
[0033] D(x1,x2)=1.5 [latitude(x2)] [0034] Westernmost distance from
node x1 to node x2 (x2 is not destination node. Each "If" statement
is tested in sequence)
[0034] AL(x2)=AdjustedLongitude(x2)=longitude(x2)-longitude(x1)
If (AL(x2)<0) then AL(x2)=AL(x2)+360
If (AL(x2).ltoreq.180) then AL(x2)=180-AL(x2)
If (AL(x2)>180) then AL(x2)=540-AL(x2)
If (AL(x2)>0) then AL(x2)=360-AL(x2)
AL(x2)=AL(x2)+1
D(x1,x2)=1.5 [AL(x2)/4] [0035] Any direction distance from node x1
to node x2 (x2 is destination node)
[0035] D(x1,x2)=1
[0036] Latitudes are normalized with 0 degrees at the South Pole,
90 degrees at the Equator and 180 degrees at the North Pole.
Longitudes are normalized with 0 degrees at the Greenwich Meridian
and increase bearing east, reaching 360 degrees upon reaching the
Greenwich Meridian again.
[0037] After determining the link most closely aligned with the
compass direction selected in step 505, routing processor 310 moves
to the next node and removes the link from the pool of available
links in network topology 305, as shown in step 525. If the
destination node received in step 405 is not the current node,
(step 530, No), routing processor 310 repeats steps 515 to 530
until the destination node is reached. Upon reaching the
destination node, routing processor 310 has found the first path
and stores it in memory device 325, as shown by step 535. Routing
processor 310 may then repeat steps 510 through 535, but in each
subsequent iteration, all the links used in the previously found
paths have been removed from the pool of available links.
[0038] Once no more available links are available (step 520, No),
all possible paths in the selected direction have been found and
routing processor 310 continues to step 540. In step 540, routing
processor 320 may rank all paths found in the selected compass
direction, for example, using the performance parameters received
in step 410. In certain embodiments, the path that optimizes the
selected performance parameters is the primary path for the
selected compass direction, and the other ranked paths are assigned
as backup paths. For example, the path with the fastest travel time
or greatest bandwidth may be chosen. After determining all possible
paths for the selected compass direction, routing processor 320 may
select the next compass direction, and make all links in the
network available to the pool of available links. Routing processor
320 may then repeat steps 510 through 545 for the new compass
direction.
[0039] Once routing processor 320 finds the possible paths for each
compass direction, in step 550 it determines which compass
directions produced at least the desired number of paths. The
desired number of paths may be one of routing parameters 315
received in step 410. If more than one compass direction has the
desired number of paths (step 550, Yes) routing processor 320 may
rank the paths for those directions by the performance parameters,
as shown in step 560. If none of the directions have the desired
number of paths, routing processor 320 may rank the paths found for
each direction in step 560 by the performance parameters. The
performance parameter may also be one of the routing parameters
315. Routing processor 320 stores all paths from the compass
direction that optimizes the selected performance parameter in
primary and alternate paths memory storage device 325, as shown in
step 565. If only one compass direction has the desired number of
paths, routing processor 320 stores the paths from that direction
in the primary and alternate paths memory storage device 325, as
shown in step 555. The paths found in the found paths memory
storage device 325 are then available for step 430.
[0040] FIGS. 6a-6n are exemplary network topologies depicting
directionalized routing processes consistent with the embodiments
of the present invention. An example of the operation of the
directionalized routing method of the present invention follows,
using the network topology shown in FIG. 6a. For the first example
assume that three link diverse paths are desired for source node C
and destination node Q. Initially, the routing processor attempts
to find three link diverse paths using a shortest path algorithm
and an iterative process in which the links used in previously
discovered paths are removed from the network--this is a standard,
well-known process returning only two paths between C and Q.
Because it has the smallest number of links, the primary path is
C.fwdarw.F.fwdarw.I.fwdarw.J.fwdarw.N.fwdarw.S.fwdarw.Q, as shown
in FIG. 6b. The routing processor removes the links used in the
primary path from the network and runs the standard shortest path
algorithm a second time. This time the routing processor finds the
backup path
C.fwdarw.D.fwdarw.E.fwdarw.H.fwdarw.L.fwdarw.M.fwdarw.O.fwdarw.Q,
as shown in FIG. 6c. Removing the links used in this backup path
separates the network, as shown in FIG. 6d, and additional paths
from C to Q cannot be calculated by the routing processor. Thus the
traditional method has failed to provide the desired number of
paths for this source and destination pair.
[0041] Compare this to an embodiment consistent with the present
invention in which the routing processor uses the directionalized
routing method by establishing the northernmost route. Starting at
source node C the path is forced in the northernmost direction, as
shown in FIG. 6e. Node D is the northernmost node connected to node
C, and is selected as the second node in the path. Likewise, node A
is the northernmost node connected to node D, and is selected as
the second node in the path from C to Q. The routing processor
follows this process until an interim or tandem node connected to
the destination node Q is reached--in this example, the first
tandem node connected to Q is node O. Once a tandem node connected
to the destination node is reached, the path is forced to the
destination node regardless of whether the destination node is the
northernmost node connected to the last tandem node in the
source-to-destination path. In this example, Q happens to be the
northernmost node connected to O. Therefore,
C.hoarfrost.D.fwdarw.A.fwdarw.G.fwdarw.L.fwdarw.M.fwdarw.O.fwdarw.Q
is the northernmost path from C to Q calculated by the routing
processor using the directionalized routing method, as shown in
FIG. 6e.
[0042] After establishing the northernmost path as the primary path
from C to Q, the routing processor removes the links used in the
northernmost path and calculates the next northernmost path, if
possible. The next northernmost path in this example is
C.fwdarw.B.fwdarw.D.fwdarw.E.fwdarw.H.fwdarw.I.fwdarw.J.fwdarw.N.fwdarw.S-
.fwdarw.Q, as shown in FIG. 6f. After removing links used in both
the first and second northernmost paths, the routing processor can
calculate a third path,
C.fwdarw.F.fwdarw.K.fwdarw.U.fwdarw.P.fwdarw.S.fwdarw.R.fwdarw.Q,
as shown in FIG. 6g. After removing the links used in the three
northernmost paths from C to Q, C and Q are both separated from all
other nodes, as shown in FIG. 6h. Therefore, no additional paths
from C to Q can be calculated and the routing processor is finished
with the northernmost path calculation. The routing processor then
ranks paths found by forcing routes to the north using a selected
performance parameter, such as delay, availability, or cost. The
path optimizing the selected performance parameter becomes the
primary northernmost path. The remaining paths are the northernmost
backup paths. Thus, for example, the primary path provides the best
performance, and the first backup path provides the second best
performance, and so on.
[0043] Once the northernmost paths are found, the routing processor
investigates the easternmost, southernmost and westernmost
directions using the method described above. Once the paths for all
directions have been found, if more than one direction returns
three link diverse paths, the direction that optimizes the selected
performance parameter (e.g., minimize delay, maximize availability,
etc) is chosen by the routing processor.
[0044] It should be noted that the routing processor uses path
efficiency when calculating the directional paths. This can be seen
in another example using the network topology of 6i. In this
example three link diverse paths are desired from source node K to
destination node A. Once again the standard, iterative shortest
path algorithm returns only two paths between K and A. The primary
path of K.fwdarw.F.fwdarw.E.fwdarw.D.fwdarw.A, as shown in FIG. 6i,
and the backup path of
K.fwdarw.J.fwdarw.I.fwdarw.H.fwdarw.G.fwdarw.A, as shown in FIG.
6j. Removing the links used in both the primary and backup paths,
as shown in FIG. 6k, separates the network and the routing
processor cannot calculate additional paths from K to A.
[0045] This may be compared to an embodiment consistent with the
invention where the routing processor applies the directionalized
routing method to discover the route between K and A. Forcing
routes to the north, the routing processor returns only two paths
between K and A. In fact, forcing routes to the north returns the
same two paths returned by the standard shortest path algorithm
described above. Next, the routing processor forces routes to the
east. The easternmost path of
K.fwdarw.U.fwdarw.P.fwdarw.S--.fwdarw.O.fwdarw.M.fwdarw.L.fwdarw.G.fwdarw-
.A is shown in FIG. 61. Note in FIG. 61, the true easternmost paths
U.fwdarw.T.fwdarw.P and S.fwdarw.Q.fwdarw.R.fwdarw.O are replaced
by U.fwdarw.P and S.fwdarw.O, respectively, by the directionalized
routing algorithm to improve the path's efficiency.
[0046] The routing processor removes the links used in the
easternmost path from the pool of available links and calculates
the next easternmost path
K.fwdarw.J.fwdarw.I.fwdarw.H.fwdarw.E.fwdarw.D.fwdarw.A, as shown
in FIG. 6m. Note that the true easternmost route of
E.fwdarw.F.fwdarw.C.fwdarw.D is replaced by the more efficient
route E.fwdarw.D. After removing links used in the easternmost
paths already found, the routing processor can calculate a third
path, K .fwdarw.F.fwdarw.C.fwdarw.B.fwdarw.A, as shown in FIG. 6n.
Thus the routing processor calculates three diverse paths between
source node K and destination node A using the forced
directionalized routing method. These paths may be ranked using the
performance criteria to determine the primary and backup paths.
[0047] FIG. 6g also demonstrates an example of path optimization.
In this figure, the northernmost route of
K.fwdarw.J.fwdarw.T.fwdarw.P is replaced by the more efficient
route K.fwdarw.U.fwdarw.P.
[0048] One of ordinary skill in the art would recognize that the
invention is not limited to the four primary compass points.
Although the four compass points will be sufficient for most
networks, all compass directions may be utilized in the present
invention for networks with high connectivity.
[0049] Other embodiments of the invention will be apparent to those
skilled in the art from consideration of the specification and
practice of the invention disclosed herein. It is intended that the
specification and examples be considered as exemplary only, with a
true scope and spirit of the invention being indicated by the
following claims.
* * * * *